top of page

Trading the Financial Markets: “Humans vs Robots, who Wins?”

“You can understand a computer in that simple way: Garbage in, garbage out…”



Several years ago back in February 1996, Garry Kimovich Kasparov (who became the record holder of the youngest undisputed chess world champion at the age of 22, when he beat Anatoli Karpov in 1985, and still -according to many experts- considered the “best player who ever walked on planet Earth”), was the first world chess champion who lost -in a highly publicized- game against a computer. The computer model was called “Deep Blue” and it was produced from IBM, especially for these series of matches. Twenty-four years ago, since this day, an Artificial Intelligence entity has beaten a superb human mind. Well, guess, what? Financial markets’ profitable trading is not a chess game, it’s far easier than that…


The introduction of today’s blog post may “sound” blasphemous, however, I can prove that at least I am saying the truth. First thing first: Trading a financial instrument in a financial market, it’s like a poker game. Has nothing to do with the trading of goods and services in the real economy. Financial markets do not produce anything they just re-distribute the already been produced and what may be produced. In simple words, their function serves not to increase the GDP of a country (or countries) or the income of a household or a firm, but to re-shuffle the “cards” and re-allocate the total money traded in different pockets than before.

Since, I believe, I’ve made my point clear until now, I will proceed with the aforementioned proof: A Chess game is a game between two participants. Either physical or artificial. Trading is a game between millions of participants, in every level of knowledge and wealth, magnitude and experience, and risk vs reward appetites. So, what makes it easier to win in this game than chess, since it seems far more complex? Well, exactly because we have many-many players in that game! And most of them are inexperienced, with mediocre knowledge or none, with unrealistic expectations, and most importantly…sentimental and therefore not rational.

So, an entity like a computer Robot, in a form of an algorithmic strategy for “playing” the trading game, has the opportunity to defeat not the world’s champion, but the average Joe who is greedy, fearful, moody and most of all…a human being. Imagine now the following experiment: we gather 100 different traders that each has a bankroll of $1000. That fact makes the total amount that will be traded one hundred thousand USD. The experiment starts with all the traders trying to make money out of price fluctuations which the real world’s changes in demand and supply of a certain tradable, creates. It is obvious that even if it’s a “50-50” game (the price of a tradable e.g. the EURUSD exchange rate, either can go up or down, in a given period), we will not have a 50-50 outcome! The most informed traders will end up getting the money of the less capable “players”. I know it seems a bit harsh to talk about trading like a “game”, but hear me now, it is just that, nothing more or less. A game of probability numbers.

The financial markets through the game of trading redistributed the previous total income of the 100 traders, to the ones who had a better strategy, comparative to the rest. Because after we repeated the above experiment over 1000 times, “luck” disappears from the horizon, and the true winners of the game, are the ones who were not counting in luck, in other words, the ones who were not gambling.

Until this moment and based on the above written, what do you think its theoretically easier? To beat Garry Kasparov or to win the bankroll of let’s say, 10 inexperienced traders? Because using simple arithmetic, to double your initial $1000, you just need to win the initial stake of another player/trader. Sounds simple? It does, and that’s what it is.

Now, who do you think, have most chances to beat a trader? Another human trader or a machine that has learned how to trade? I guess the answer is pretty obvious: it depends on who and in what way taught the machine to trade. But given the same strategy, the machine has much more probabilities to win over a human, just because it doesn’t get tired or emotional. The machine doesn’t have any feelings, namely no fear, no greed, no hesitation to pull the trigger, no need for sleep, or having a bad mood because of a late previous night’s quarrel with its wife. An algorithm doesn’t get sick. An algorithm doesn’t get disappointed nor over-enthusiastic. Data is all that it needs to be fed so to function.

I firmly believe that I have proved that is not only easier for a computer program to make money out of financial markets by beating other less capable human traders or less capable other machines, because of the number of the participants and because of the characteristics of the game itself, but also I have proved that an average human being has practically slim or no chances to win over an average algorithm. And I remind you, that I am not talking about ONE trade. I am talking about over 1000 trades. Someone may say now, “wow, 1000 trades, that’s 100 trades per year so we need 10 years to make them”. Well, it’s a decent argument but given the ability of a computer program to interact in fractions of milliseconds, and to execute its decision-making algorithm, maybe 2 months are enough. I know of algorithms that commit 1000 trades per hour…but they are practically out of reach for us the retail traders. This belongs to the HTF (High-Frequency Trading) category, that is not a preferable way to trade the markets, at least for me. Hence, kinds of HTF algorithmic trading utilizing special order placement/replacement algorithms such as Predatory trading, Pinging, Point of Presence, or Liquidity Rebates are not and will not be in our horizon of interests.

When you decide to risk your hard-earned money, the first thing you need to do is to realize what kind of trader you are. And to do that you have to consider not only your bankroll but also be self-aware of your knowledge-based level. Well, guess what, again: If and I say just “if”, you understand that to be a complete trader it might take years and a great cost in personal time, if and I just say “if”, you understand that a machine, a computer program, an algorithm can do that for you now, while you sleep, or work, or while you share some time with your loving people, you may take the smart decision to hire a robot under your employment. To hire a computer that will do the hard work for you.

In any case and speaking very frankly, you don’t stand a chance if you decide to do that alone. Not if you won’t be able to sacrifice the next two to five years of your life, to become a good trader. What would be your chance to win in a chess game against Kasparov or any other previous or later word champion? Against let’s say, my favorite chess player, the late great legendary Bobby Fischer? The answer is “Slim or None”. But, just think about it, what if you were the owner of IBM’s “Deep Blue”? Then your probabilities would have exploded exponentially higher.

Finally, and returning to the question that the title of this blog postposed, namely who wins in the trading game a human or a robot, I believe the answer has been given:

The ultimate winner will be a human who has the best robot.

17 views0 comments
bottom of page